Duration 33:15
16+
Play
Video

Build for Android (Go edition): optimize your app for global markets

Raj Ajrawat
Global Product Specialist Lead at Google
+ 3 speakers
  • Video
  • Table of contents
  • Video
2018 Google I/O
May 8, 2018, Mountain View, USA
2018 Google I/O
Video
Build for Android (Go edition): optimize your app for global markets
Available
In cart
Free
Free
Free
Free
Free
Free
Add to favorites
6.96 K
I like 0
I dislike 0
Available
In cart
Free
Free
Free
Free
Free
Free
  • Description
  • Transcript
  • Discussion

About speakers

Raj Ajrawat
Global Product Specialist Lead at Google
James Bender
Product Manager at Google
Amrit Sanjeev
Developer Advocate at Google
Steve Suppe
Product Manager at Google

Raj is the Product Specialist for the Android Platform on Google Play, covering products such as Android Instant Apps, Android Oreo (Go edition), Android platform APIs and Android OS releases. He earned a bachelor's degree in Economic Policy from New York University, where he graduated magna cum laude and was a recipient of the NYU Founder's Day Award for academic achievement.

View the profile

James is a product manager on Google Play concentrating on all things emerging markets. Prior to Google, James was a US Navy Pilot with more than 400 aircraft carrier landings.

View the profile

Amrit Sanjeev works as a developer advocate in the partner developer relations team and works with partners in India. Previously he managed multiple programs like Google developer communities (GDG) , Google developer expert (GDE) and developer agencies program. Prior to joining Google he worked for companies including Intuit, Philips and IBM. He was the first Android GDE from India and actively participates in developer communities in India.

View the profile

Steve is the serving and distribution product manager on the Google Play Console. Steve is based in London, grew up in Southern California, and has lived in Colorado, Mississippi, Northern California, New York, Israel and Switzerland. He's a Fulbright Scholar in Computer Science and holds an MBA from London Business School. Steve's current favorite album is Live Rhymin' by Paul Simon.

View the profile

About the talk

With the launch of Android Oreo (Go edition), anyone can get a great Android experience on entry-level hardware. Learn about optimizations that will help you grow a global audience in countries like the US, Brazil, India, Indonesia, Mexico, Nigeria, and many others. You'll get insights from developers who have seen results by applying recommended best practices. You'll also get an introduction to new ways Google Play is adapting to the challenges people face installing and keeping apps updated.

Share

Hey everyone. My name is Rodger drop it and I'm a product specialist for the Android platform here at Google play today. I'm going to begin our session by talking about the Android GO opportunity and walking through some details about the product since we last spoke about it. I'm thinking it's like my colleague Steve to come up on stage and walk you through some of the data that we're seeing support all the details as to why we build Go including reasons why you should keep your app small. After that, we'll have Elmer's come up on stage 2 walkthrough all the requirements that you

should pay attention to when it comes to optimizing your app and your game for these devices. And finally I'm going to have James come up on stage and he's going to walk you through all the details that we've been doing on the Play Store to make sure that we're delivering a really awesome store experience for users on these devices. To start things off Android go with an ecosystem wide effort to make sure that we delivered in premium hardware and software experience for users on entry-level devices. Now in all of the user research that we did globally

we found that users had very specific pain points on these types of devices those items included processor speed and quality storage capacity and battery life. No matter where we went in the world in developed markets in developing markets users consistently, coming up with those issues on these types of devices ourselves along with our OEM Partners to deliver a much more integrated and better experience at this device capability. It's important to remember that Android go is not a fork of Android. What we did was we actually introduced the

bill time flag into the platform that it ensures that all devices that are 1GB of lesser Ram going forward are going to be go configuration devices. We work with all of our Google teams internally from search to the assistant to YouTube and maps to make sure that we delivered a first-class Google experience as well. So users are on these devices. We also made sure to introduce preload limits. We don't we don't want to have users unboxing super slow phones II they buy them a little bit more into storage performance and security. Now again, referencing that

user research that we were talking about earlier. We noticed that over two-thirds of users on devices with 8 gigabytes of flash or less really only had about a gigabyte of storage when it came down to it. You can't really do much with a gig of space. It's 2018. I want to take pictures. I want to download apps. I want to play games at what we realized was we need to do a better job of actually making sure that users had more storage out of the gate and we'll work will be when we were able to accomplish with software improvements was that users now have 2X more storage right off the bat

that's an additional photos that they can take on their phone. Now it would be disingenuous of me to come up here and ask all of you to optimize your apps in your games for these devices if we didn't ourselves actually do the same thing. So we got together with all the different products teams for some of the major apps that you know and love like search the assistant Gmail YouTube and make sure that we delivered a really good experience from Google on these particular device has some of our product team decided to remove some features. I'll just decided to add specific

features for these two types of devices. It really just depended on the use case for those particular team's one of our most popular apps is actually files go and in that case scenario, we realize that we needed to deliver a better job for users in terms of managing space on their phones. We've been really excited to see some of the the the the ratings of the excitement that users have with these particular apps. Now security is at the heart of everything that we do at Android and one of the things that we launched as part of this was Google Play protect with the

rise of peer-to-peer sharing and end offline usage of apps. We wanted to make sure that users felt very comfortable with the apps and the games that they were downloading onto their phone think of play protect as a badge of trust that essentially extended the play Brand to let it user know that the app for the game that they're putting on their phone can be trusted to be used on that particular phone. My colleague James will go through more details about what that actually looks like on a user experience for these devices. Now I'm really proud of all the work that we've accomplished so

far over the past year, but let's take a second to dive little bit more into the opportunity around why it matters for you to optimize your specific experience for these types of devices. This chart right here shows 2017 shipments globally for Android devices broken down by the the RAM on those devices as you'll see on the top right there about 25% of global shipments. Every year are devices that have 1GB of lessor ram 1 gigabyte of Ram or less that represents about 300 million devices

in the market currently. We expect the number to actually increase throughout the year, but it just kind of gives you a general sense for how many devices you should expect will actually have this configuration going forward. We seen a ton of interest in the market from all of our OEM Partners. In fact back at Mobile World Congress in March. We had about six OEM launch devices. I'm happy to report now we have over 80 m a t o e m Partners committed to delivering devices with hundreds more in the pipeline by the end of the year. We really cannot thank our partners enough for

all the work. They've done to partner with us to deliver a really awesome Hardware experience for entry-level devices encourage you to go out and buy a device so you can feel and see just how premium the experience is at such a decent price point. Now I want to dive a little bit more into the Indian market specifically India represents the largest market for these types of devices. We expect about 30 to 35 million devices by the end of this year. I don't have to go configuration on them. Marketing that you're seeing right here is what Micromax

is actually using in Market to show users. Just how premium the experience is end up going to call out is that with carrier subsidies many of these devices are actually under 40 US dollars in these markets. So very very good price point and still premium Quality Hardware. Of course India does represent the largest opportunity for these devices, but I can't stress this point enough. The u.s. Is the second largest market for these devices today. Every time I tell that to people everyone looks at me with blank stares because it doesn't make any sense the numbers prove

out that there's the news as the second-largest market for entry-level devices today. This isn't just a developing Market problem or an emerging-market problem. This is a develop Western Market First World Market problem as well users around the world at these price points and at these device configurations want to have a premium experience and go delivers that I'd like to know invite my colleague Steve to come up on stage to walk you through some of the data that we're seeing on the play side around all the information that we see that supports the reasons why you should invest

in the opportunity as well. Thanks Ross. Hi everyone. My name is Steve. I'm a product manager on the Google Play console just introduced you to answer I'd go and talked a little bit about all the work that we've been doing to shrink the storage and memory footprint about Google apps in preparation for it. I want to talk to you a little bit about the importance of doing your bit and the rewards you'll get to your business and your user Relationships by focusing on shrinking your app size and

thinking small. Now if you haven't gotten enough data analysis over the last few days will want your hats going to show you a bunch of grass. Add devices have gotten more and more advanced developers like you are making increasing use of their capabilities, which is great and means you're packing in it more features than ever before you're making better user experiences providing providing more value to users, but I can come at a serious cost this chart here that I'm showing and you might have seen her earlier on Google Play over the last 10 years

and you can see very clearly the app size of continue to increase. In fact, it's going to 150% over the last five years run and it's still growing all the increasing size may be materially impacting your user acquisition strategy. App size has a very real effect on conversion rates and a successful download of your at this chart shows a correlation between app size and user conversion rates specifically what I'm talking about conversions. I'm talking about users who come to the Play Store and tap install. That's the conversion or talking about here

clearly the heavier your app. You can see app size as the x axis the less likely a user will be to install your app. Even when we account for things like store listing optimization size has a very clear negative impact on whether or not you want to install your app on their device. In fact data shows that for every 6 megabytes that increases on a nap size. We see a decrease in the user conversion rate of roughly 1% to all apps under a hundred and also applies Global a conversion that could be recovered if you focus on

that size, but that's not all we can also compare country-by-country. Now if we compare that percentage increase in conversion by reducing app size by 10 megabytes, we first of all sea by the global average that worldwide users see more benefit they tend to install more But this can vary dramatically country-by-country as well shrinking around by 10 megabytes in a country like India or Brazil has a larger impact on his door conversion rate than countries like Germany the USA and Japan, but if you look across all Emerging Markets the removal of 10 megabytes from a

nap size roughly correlates with an increase in conversion rate by about 2.5% Let's drill down a little bit farther when we talk about business as we talked about conversion rates, but we have a different lens that you might not have seen before because we're also the ones that serve the app. So what happened after I use their tops install. Well, they still have to download the app, right so we can also measure what percentage of apps by size actually fully make it to the device to have 100% transfer rate and are ready for install. That's what

this graph shows here. So you can see that as app size increases the success of a downloaded drops precipitously. This could be for a variety of reasons that aren't all shown here. So for example, I might be for environmental conditions perhaps in certain parts of the world Network quality isn't as high as we would like it to be as reliable as we would like it to be a might also be for a user action. I have to use a gets tired of waiting around for a large app to download or maybe once they're on the download screen. They see just how large the app. besna decide to cancel the blue line

here shows apps give me devices that have less than or equal to 1 GB of RAM, which is the same device profile is the Android GO devices were talking about and you can see the impact is even more dramatic for those sorts of devices. But whatever the individual reason you can take today is clear. Keep your app small. in many parts of the world you just pay attention to app size when deciding what to download this chart shows the median download size country-by-country for devices with less than or equal to 1 GB ever am so in countries with cheap data or cheap

storage or strong Network reliability the blue But in many markets particularly across Africa and Asia, we see a clear Trend towards smaller apps. But the three-headed reiterate Rogers point this isn't just an emerging-market problem. If you look at that Medium download size again, we can see that the u.s. Is roughly right in the middle of our Global minimum and maximum size that roughly 50% of users in the US and countries like it also are more likely to prefer smaller at

the u.s. Is the second largest market for Android go devices globally and I think he's number sort of helped to support that statement. Remind me thanks yourself. What's the point of this part of the presentation is Google really trying to tell me to remove features for my app and and try and make my apps tomorrow by doing that not exactly. What I'm trying to do is paint for you at a larger picture of the decisions in the factors that you should weigh. When considering what to do in your ass next some of your decisions might might cause you to lose out on quality install prevent users

from installing your app Optimizer app for Android go Banksy my messages, Google based on India. I want to talk about how you can optimize your apps to make them Android go ready and working with our partners is that when they actually optimize their apps for Android go it just did not benefit only that users who use Android Wear devices. It did help everyone who was using those do apps work with people who are in Emerging Markets as well as using high-end devices. I can hopefully see during these apps with these requirements will

enable all your users and will bring a lot of benefits for you. Firstly you need to set your target SDK 226 or about this is not only ensure that the latest Android experience to users. It will also allow you to comply with a play policy change that requires you to set the target SDK 2 and -1 API level 10 is the latest Android API. Even getting abortion lower right now. You need to ensure that you cast your ass. Early to make sure that there are no brake adjust or functionality

changes because of this API change their behavior changes in eBay 26 and you need to account for that. Limitations like background processing location updates notifications are some of the examples of reasons why you might want to test motor relay. The second requirement is release date biggest size for apps. Wheelfire to make it less than 40 m b and her games as extender to 65 MB. That's excluding the secondary downloads to make it as small as possible. Cancel all the things that you could do in

order regular APK size remove all unnecessary code on a nursery library in adding a mobile-optimized and not server-side one. The drawl doesn't a resourceful. They need to be optimized for both sides as well as the format. Even if it'll give a backward compatibility and adding support me for Mario's and show that you're adding only the models that are required by the application and not all of them in the audio source files like. Wav files to MP3 or AAC with a much smaller format. So there by reducing the size

you need you can use the Android APK and lies all that shipped along with Android Studio to check the sizes of the APK APK against the previous one as you make changes to see how well you're making increments and US1. Either using the new app windows for Mac. Then you might want to use the bundle tool and check the size of the celestial. We all know that entry-level devices are I'm constrained. So you need to check all your key user features to ensure that they run properly without

any issues on a fight with ambalamb device. E&R Jackie frames out of memory examples of things that you go to watch out for me before you leave them for these problems and fix them proactively before you send it out in the market once you actually put it out there in the market you need to observe and we actively fix issues as they come and red white blue which James will actually talk about later on Franklin and 5s a performance monitoring tool are some of the obstacles that you could use in order to do this reactive fox really really well.

Let's take a look at the requirement for RAM usage. Stop remembering your ass should be less than 50mb and 4 games that extended 150 MB. I want to make clear that what I'm talking about RAM usage and talk about the proportional stats, Isaiah. Rashad is basically the sum of the private memory or after using plus the proportion of the shared memory with other processors. That's all you can use a Journeys and optimize them for memory usage and make the memory footprint as low as possible. A common way to find SPSS

number is to use the mmmm Pokemon Db shell memory loss and programmatically Camp Clippers. Let's move on tap start a timer app for stock in less than 5 seconds in 5 seconds, but I mean is beef fully interactive in less than five seconds. Quick start a key to avoid early on installs and abandonment. A quick start of time. Is he in building a perception that your app loads fast. And as far as for the user as a general rule you need to ensure that is a plus perceived to be fast on these devices. Another measure of the apps. Of time you could use the finest Performance Marketing SDK, which

athlete athlete is automatically chosen on the fireplace console for you. We're extremely proud of the work of artists have done in order to reduce size and improve performance for there. For the users. We can't thank them enough of partnering with us on this initiative and providing a fast performance. Enriched experience for the users are now show you a video and Athena's put together. I need the experience Flipkart team had while they were optimizing that are for Android. Let's take a look.

The car is India's leading e-commerce player. Its mission statement is always being how do we enhance the living experience the lifestyle of finding users by bringing the most affordable and most quality products to all sections of the country using technology. One of the biggest learning from our user base was lot of these devices were actually with less than 1 GB of RAM where people were actually aborting the journey because the phone could not handle the storage or the competition requirements We Don't Starve Android go application. If I can load my pages

Foster, I can give those additional milliseconds back to the user which can actually give the user a much better experience in the shopping Journey on foot cut the app install size as Birds Go Android go guidelines since to be less than 40 MB and with some good guidance from the Google team. We have brought that down. We are using Android studio recital and proguard service in helpful for reducing resources and Cool Base. The guideline the free received was the proportional set size should be less than 50 mb.

15.5 screen reduce our with my phone cases from fullscreen. The one is clean working on the Android app for credit card has become much lighter. It has begun will perform existing user best of actually seeing the uninstall rates go down by 44%. Versions now crash almost ten times less as compared to our older versions is an average rating of 4.4 plus a lot of effort has gone behind reaching back Milestone get excited with Android go levels be crossing the 200 million very very

easy and very very soon because we believe the next set of online users are actually going to come on devices for which Android go is the right medium to approach. So if you're just go get out understand your user try to make you Technology Taylor 2008 and then you would have an app which would be loved by your users. This so proud of the work Flipkart and all apartments have done in order to optimize the last for Android go. I would like to know how I'm going to

handle this case to my colleague James who's going to talk with you. Talk to you about the play updates. Thank you. Thank you. Everyone, I'm James and I'm a product manager on Google Play. I work on initiatives like Android go emerging markets and end-user optimization, and I'm really excited to share all the work that the entire play team has put in over the past couple of years to build or quality Android experience to our users or I'm going to chat a little bit of experience and then optimizations of improvements. We made under the hood and then I'll tell you

guys about some of the tools that we've developed for you to make it easier to build on Android go. A key point that I want to make today is that there is no Go version of the Play Store. It's just the play store optimized for go the same Play Store on every device. We did this deliberately because we think that the optimizations that we can make should be the customer should be able to take advantages of those optimizations on any device that they may be on will there be a go device or pixel. We aren't living in the catalog on Android go so users get access to any title

that you choose to distribute to them whether your distribution settings be limited by location or by device type. So we kind of need to set the example you've heard today over and over the small matters. We went out over the past year and have more than the Play Store and ask for massage perspective and we're now just at or below 10mb on all devices globally. It's in a couple of super proud of users in Emerging Markets have made it Ultra clear to us that mobile data. Is there currency so

we ran a couple of tests and we showed how expensive any app download is on a card and on search results and as you can imagine for a smaller apps, even if a rating maybe lower so in a couple of places, we actually replaced where we previously had an APK an operating with an app size. We found that showing APK size increases install rates and more importantly user engagement overtime. Remember what Steve said earlier APK size does have an effect on install conversion and successful install rate.

and with the with the rise of offline and peer-to-peer sharing play wanted to ensure they use it could still trust the apps and games that they got from us. No matter where those apps came from. That's why we play protect you can think of it as a badge of security that users will know and trust So what this does is it gives users device gives users on any device type confidence that there have a safe and secure app experience when they're in interacting with these apps in the context of play. How do we do it? We wanted to make sure that an app to be trusted on

users device. So we started to attach a very small bit of security metadata to every app and game that we publish then what we can do is if user requires an app offline or via peer-to-peer we can use play protect to Detroit to read that security metadata and determine and apps authenticity and verified Source you as a developer get the added benefit of being able to service those apps because we will bring them into the user's Play library. And will be able to update those apps in the future some major pain point that we know you've relayed to us. We're excited to

address moving forward. APK size the downstream metrics that come from it installs how long a user may keep an app on their device. And APK size getting bigger over the past 10 years cause us to really take a step back and evaluate ourselves. And we said what can we do to make this situation better? Well through that analysis we came up with Google Play Dynamic delivery today. You can use Android app. Android app bundle to benefit from a smaller app without needing to refactor your code the app bundle

also enables future modularization. What it does is it reduces your app at the moment the user installs the app and allows you to load Dynamic feature modules on demand at the point when they're needed further down the road when the user is engaging with the app rather than making sure there's lots of pressure out of salt. I'm the Android app on the along with Google Play Dynamic app delivery ensure. We deliver apps to the Future that are modular instant and dynamic. We've also tuned the Play Store's personalization algorithm to show smaller apps on lower-level

devices and as he could imagine this increases engagement because downloads are more successful and smaller apps run better on these entry-level devices. This is evident and all the data we've collected on user behavior and its key to think about how size will affect App Store ranking for these users. We've done a bunch of a feature improvements for the store. For example, when a device or when the user is near device disc capacity space. We suggest users removing apps that have either never been opened or haven't been used in a long time. You can see on the

screen here. We show a mock apps that are of a specific size that haven't been used from least frequently used most frequently used and we give the user who is trying to install you're at and opportunity to delete an app that its tail or they or they never used that gives them an opportunity to have a better experience with your app. Another thing that we recognize we were not serving our users in the best way was the notion that a lot of users manage app installs when they're on

Wi-Fi. Well, we thought we could do better. So we created plays wait for Wi-Fi feature. It's a known usage pattern everyone sees all the time, but we thought play could do a better job of serving a particular user need. Data management is important to these users. So if today if a user is on play Via Mobile connection, they can either when they hit install they can either choose to install that app immediately or delay that download and tell their device connect to Wi-Fi and will initiate the download in the background and alert them when they when the download and install is

complete. This gives users more control. It says metered data and most importantly it saves money. We also took a look ourselves and realize that we weren't doing a very good job from an offline or a connectivity tolerance for interview. So we built a couple of additional features. The first one is when the Play Store is on Wi-Fi. We cash a larger percentage of the most visited parts of the store. So what that means is if user turns mobile data off or their connectivity is crummy or their connectivity is spotty. We can actually serve a greater percentage of the store

directly from cash. This is great because it works offline. It's also great because it's a super-fast experience for our users then when they reach the edge of the cash there offline and they can't bring up any additional Pages we built another feature that allows the user to opt into a notification that will bring them back to that part of the store that they left once their device establishes a connection again. This is a great opportunity. But user to get back to play to download your app. All the features. I just walked through are designed to give end-users a high-quality

store experience. But I also want to reiterate that we have tools available for you to help you optimize your app for the Android GO devices and these markets the first which which no one in this room should be unfamiliar with is Android vitals. Android vitals gives you an insight into and ours and crashes and it's in Paramount for understanding stability of your app battery. And the most important part is that Android vitals performance in packs or promotability on the Play Store, especially in the context the search and discovery. We also introduce

a flag. ASRock previously spoke of called is loram that allows you adjust your apps ux and distribution choices. We want you to optimize for maximum distribution Pearson additional tool to help with that. We've also included details about Android go devices in the console pre-launch report. This allows you to check on your builds understand how they're performing on actual Android go devices on test bad before you release the market be sure you incorporate a regular Cadence with the console pre-launch report as part of your normal keyway and release

management process. So you guys have been to a lot of meetings you've heard a lot of things and if you've been on your phone the last 25 minutes before things you need to remember Android go is a huge opportunity. Keep your off small play is doing everything we can to improve after excuse me. You should do everything you can to improve app quality by levying leveraging the Android tools and Palais plays delivering a focused experience for these users. Thank you everyone. If you have any questions, the four of us will be up here at the front. Looking forward to hearing

from you.

Cackle comments for the website

Buy this talk

Access to the talk “Build for Android (Go edition): optimize your app for global markets”
Available
In cart
Free
Free
Free
Free
Free
Free

Access to all the recordings of the event

Get access to all videos “2018 Google I/O”
Available
In cart
Free
Free
Free
Free
Free
Free
Ticket

Interested in topic “Software development”?

You might be interested in videos from this event

September 28, 2018
Moscow
16
159
app store, apps, development, google play, mobile, soft

Similar talks

Marc Bachinger
Developer Advocate at Google
+ 1 speaker
Andrew Lewis
Software Engineer at Google
+ 1 speaker
Available
In cart
Free
Free
Free
Free
Free
Free
Jonathan Gerrish
Software Engineer at Google
+ 1 speaker
Stefan Ramsauer
Software Engineer at Google
+ 1 speaker
Available
In cart
Free
Free
Free
Free
Free
Free
Serge Beauchamp
Software Engineer at Google
+ 3 speakers
Mike Davis
Engineering Manager at Google
+ 3 speakers
Nicholas Lativy
Software Engineer at Google
+ 3 speakers
Radha Narayan
Software Engineer at Google
+ 3 speakers
Available
In cart
Free
Free
Free
Free
Free
Free

Buy this video

Video

Access to the talk “Build for Android (Go edition): optimize your app for global markets”
Available
In cart
Free
Free
Free
Free
Free
Free

Conference Cast

With ConferenceCast.tv, you get access to our library of the world's best conference talks.

Conference Cast
558 conferences
22059 speakers
8245 hours of content